Pemrosesan Sinyal Data dan Citra Digital (PSDCD)

Topik:
Linear & Non-Linear Filtering, Fast Fourier Transformation Theory,
State Estimation, Pattern Recognition, Identification Theory

Dosen:
Dr. Ir. Andrian Andaya Lestari, M.Sc., SMIEEE
 Informasi Mata Kuliah
• Mata Kuliah : Pemrosesan Sinyal Data dan Citra Digital (PSDCD)
• Kode Mata Kuliah : T-TP 2203
• SKS : 3 SKS
• Dosen Pengampu : Letkol Laut (KH) Dr. Gentio Harsono, ST, MSi
• Topik : Linear & Non Linear Filtering, Fast Fourier Transformation Theory, State
Estimation, Pattern Recognition, Identification Theory
• Jadwal : Semester 1 – Minggu 2
• Dosen : Dr. Ir. Andrian Andaya Lestari, M.Sc., SMIEEE
• Email : [email protected]
• Tujuan : Mampu menjelaskan dan menggunakan teknik-teknik DSP sesuai topik di
atas.
• Buku Acuan : Lizhe Tan & Jean Jiang, Digital Signal Processing: Fundamentals and
Applications, 3rd Edition, Academic Press (Elsevier), 2019.
 Contents
• Introduction
• Linear & Non-Linear Filtering
• Fast Fourier Transformation (FFT) Theory
• State Estimation (tugas)
• Pattern Recognition (tugas)
• Identification Theory
 Introduction (1)
• An analog signal is continuous in both time and amplitude. Analog signals in the real
world include current, voltage, temperature, pressure, light intensity, and so on. The
digital signal contains the digital values converted from the analog signal at the
specified time instants.
• Analog-to-digital signal conversion requires an ADC unit (hardware) and a lowpass
filter attached ahead of the ADC unit to block the high-frequency components that
ADC cannot handle.
• The digital signal can be manipulated using arithmetic. The manipulations may
include digital filtering, calculation of signal frequency content, and so on.
• The digital signal can be converted back to an analog signal by sending the digital
values to DAC to produce the corresponding voltage levels and applying a smooth
filter (reconstruction filter) to the DAC voltage steps.
• DSP finds many applications in areas such as digital speech and audio, digital and
cellular telephones, automobile controls, vibration signal analysis, communications,
biomedical imaging, image/video processing, and multimedia.
 Introduction (2)
• Basic concept of Digital Signal Processing (DSP):

Sample-and-hold

• Examples: digital/Internet audio or video; digital recording; CD, DVD, MP3

players, iPhone, and iPad; digital cameras; digital and cellular telephones;
digital satellite and TV; wire and wireless networks; medical instruments:
digital electrocardiography (ECG), digital radiography, etc.
 Introduction (3)
• Digital filtering:

• The DSP block can operate as a digital

lowpass filter.
• After processing the digitized noisy signal x(n),
the digital lowpass filter produces a clean
digital signal.
 Introduction (4)
• Signal frequency (spectrum) analysis:

• Example: A digitized audio signal and its

calculated signal spectrum
(frequency content), defined as the signal
amplitude vs its corresponding frequency
obtained using a DSP algorithm,
called Fast Fourier Transform (FFT).
 Introduction (5)
• Signal frequency (spectrum) analysis:
• The top plot shows the digital speech
waveform vs. its digitized sample
number, while the bottom plot shows
the frequency content information of
speech for a range from 0 to 4000 Hz.
• We can observe that there are about 10
spectral peaks, called speech formants,
in the range between 0 and 1500 Hz.
• Those identified speech formants can be
used for applications such as speech
modeling, speech coding, speech
feature extraction for speech synthesis
and recognition, etc.
 Introduction (6)
• A picture of an outdoor scene is taken by a digital camera on a cloudy day.
Due to this weather condition, the image is improperly exposed in natural
light and comes out dark.
• The image processing technique called histogram equalization can stretch the
light intensity of an image using the digital information (pixels) to increase
the image contrast, therefore, detailed information can easily be seen in the
image.
 Linear & Non-Linear Filtering (1)
• Signal Sampling and Quantization:

• Nyquist sampling theorem:

The sampling theorem guarantees that an
analog signal can be in theory perfectly
recovered as long as the sampling rate is at
least twice of the highest-frequency
component of the analog signal to be
sampled, or
Linear & Non-Linear
Filtering (2)
• From Fig. 2.6, it is clear that the
sampled signal spectrum consists of
the scaled baseband spectrum
centered at the origin, and its
replicas centered at the frequencies
of kfs (multiples of the sampling
rate) for each of k = 1, 2, 3,...
• Anti-aliasing & anti-image filters:
 Linear & Non-Linear Filtering (3)
• Block diagram for a DSP system:
 Linear & Non-Linear Filtering (4)
• Digital Signals & Systems:
 Linear & Non-Linear Filtering (5)
• Linear, time-invariant, causal systems:

Linear systems Time-invariant system

 Linear & Non-Linear Filtering (6)
• Difference equations:
 Linear & Non-Linear Filtering (6)
• Difference equations:
Linear & Non-Linear
Filtering (7)
• System representation using its
impulse response:
 Linear & Non-Linear Filtering (8)
• Z-transform:
 Linear & Non-Linear Filtering (9)
• Z-transform:
 Linear & Non-Linear Filtering (10)
• Linear filtering: Finite Impulse Response (FIR) filters:
• An FIR filter is completely specified by the following input-output relationship:
Linear & Non-
Linear Filtering
(11)
• Linear filtering: Finite
Impulse Response
(FIR) filters:
 Linear & Non-Linear Filtering (12)
• Linear filtering: Finite
Impulse Response (FIR)
filters:
• It is a linear system.
• The stability of the filter is
guaranteed.
• Its impulse response has only
a finite number of terms.
• The FIR filter operations
involve only multiplying the
filter inputs by their
corresponding coefficients
and accumulating them.
• The implementation of this
filter type in real time is
straightforward.
 Linear & Non-Linear Filtering (13)
• Linear filtering: Finite Impulse Response (FIR) filters:
 Linear & Non-Linear Filtering (13)
• Non-linear filtering: Infinite Impulse Response (IIR) filters:
• An IIR filter is described using
the difference equation:

• The IIR filter transfer function is given by:

• The IIR filter output y(n) depends not only on the current input x(n) and past inputs
x(n-1), …, but also on the past output(s) y(n-1) …, (recursive terms). Its impulse
response has an infinite number of terms.
• Comparing with the FIR filter, the IIR filter offers a much smaller filter size. Hence, the
filter operation requires a fewer number of computations, but the linear phase is not
easily obtained. The IIR filter is preferred when a small filter size is needed but the
application does not require a linear phase.
 Fast Fourier Transformation Theory (1)
• The algorithm transforming the time
domain signal samples to the
frequency domain components is
known as the Discrete Fourier
Transform (DFT).
• The DFT also establishes a
relationship between the time
domain representation and the
frequency domain representation.
• Therefore, we can apply the DFT to
perform frequency analysis of a time
domain sequence.
• The DFT is widely used in many other
areas, including spectral analysis,
acoustics, imaging/video, audio,
instrumentation, and
communications systems.
 Fast Fourier Transformation Theory (2)
• Fourier transformation of periodic digital signal:

• According to Fourier series analysis, the coefficients of the Fourier series

expansion of the periodic signal x(t) in a complex form is

• And its digital form is

 Fast Fourier Transformation Theory (3)
• There is an important feature of the previous equation in which the Fourier
series coefficient ck is periodic of N.
 Fast Fourier Transformation Theory (4)
• Discrete Fourier Transform (DFT):
• First, we assume that the process acquires data samples from digitizing the
interested continuous signal for a duration of T0 seconds. Next, we assume
that a periodic signal x(n) is obtained by copying the acquired N data samples
with the duration of T0 to itself repetitively.
• The Discrete Fourier Transform becomes:

• The inverse Discrete Fourier Transform becomes:

Fast Fourier Transformation
Theory (5)
• Discrete Fourier Transform (DFT):
 Fast Fourier Transformation Theory (6)
• Fast Fourier Transform (FFT):
• The FFT is a very efficient algorithm in computing DFT coefficients and can
reduce a very large amount of computational complexity (multiplications).
• We consider the digital sequence x(n) consisting of 2m samples, where m is a
positive integer, that is, the number of samples of the digital sequence x(n) is
a power of 2, N = 2, 4, 8, 16, etc.
• If x(n) does not contain 2m samples, then we simply append it with zeros until
the number of the appended sequence is a power of 2. This process is called
zero padding.
• FFT algorithms: “the decimation-in-frequency algorithm” and “the
decimation-in-time algorithm”. They are referred to as the radix-2 FFT
algorithms (see the reference book).
 Fast Fourier Transformation Theory (7)
• Fast Fourier Transform (FFT):
• Zero padding does not add
basic information and does not
change the spectral shape but
gives the “interpolated
spectrum” with the reduced
frequency spacing. We can get a
better view of the two spectral
peaks described in this case.
• The only way to obtain the
detailed signal spectrum with a
fine frequency resolution is to
apply more available data
samples, that is, a longer
sequence of data.
 Fast Fourier Transformation Theory (8)
• Fast Fourier Transform (FFT): the decimation-in-frequency algorithm
• The DFT can be written as:

• Mathematical manipulations
leads to:

• It can be summarized as:

 Fast Fourier Transformation Theory (9)
• Fast Fourier Transform (FFT): the decimation-in-frequency algorithm

(200 times faster!)

 Tugas Terstruktur
• Buatlah rangkuman (summary) sepanjang max. 2 halaman A4 tentang topik:
§ Mahasiswa no urut 1 - 8: State estimation.
§ Mahasiswa no urut 9 - 17: Pattern recognition.
• Masing-masing mengirimkan ke email: [email protected] paling lambat
sebelum UTS semester ini.